Skip to content

README.md

Latest commit

 

History

History
155 lines (102 loc) · 4.66 KB

README.md

File metadata and controls

155 lines (102 loc) · 4.66 KB
Raw

Python 3.12+ UV FastAPI Docker GCP CI/CD

🎌 RekuAI — Anime Recommender System

A Scalable, Hybrid AI Recommendation System for Anime Lovers
Powered by Machine Learning, DVC, Jenkins, Docker, and GKE

License Build Status Python

"Anime meets AI. Recommending your next obsession, one episode at a time."

🌟 Project Overview RekuAI is a hybrid recommendation engine tailored for anime fans. It combines collaborative filtering and content-based filtering to deliver highly personalized anime suggestions. This system leverages a full MLOps pipeline — from fetching data stored on Google Cloud Storage (GCS), through model training and versioning with DVC, automated CI/CD using Jenkins DinD, all the way to deployment on Google Kubernetes Engine (GKE).

🚀 Built for scale, speed, and smarter recommendations.

🧠 Recommendation Engine RekuAI uses a hybrid recommendation system:

Component Technique Description 🎯 Content-Based Filtering TF-IDF + Cosine Similarity Recommends based on anime genres, synopsis 🧑‍🤝‍🧑 Collaborative Filtering Matrix Factorization / SVD Learns user preferences from ratings ⚙️ Hybrid Engine Weighted Blending Combines both approaches for best results

📦 Key Features

  • ✅ End-to-end ML lifecycle orchestration
  • ✅ Data ingestion from Google Cloud Storage (GCS)
  • ✅ CI/CD pipeline using Jenkins with Docker-in-Docker setup
  • ✅ Model API built with FastAPI and containerized
  • ✅ Cloud deployment with Kubernetes Engine Clusters

📁 Folder Structure

RekuAI/
├── src/                    # Core logic: recommenders, data pipelines
├── pipeline/               # ML pipeline orchestration
├── notebook/               # Jupyter notebooks for exploration
├── utils/                  # Helper scripts
├── templates/              # HTML templates (Flask UI)
├── static/css/             # Styling
├── .dvc/                   # DVC metadata
├── Jenkinsfile             # Jenkins pipeline config
├── Dockerfile              # Docker image build
├── deployment.yaml         # Kubernetes deployment spec
├── application.py          # FastAPI app entry point
├── .env                    # Environment variables
└── README.md               # You're here!

🚀 Getting Started

🔧 Prerequisites

Python 3.8+

Docker

DVC

Jenkins DinD setup

Access to GCP (GCS & GKE)

kubectl & gcloud CLI tools

🧪 Setup Instructions

bash Copy Edit

1. Clone the repo

git clone https://github.com/yogeshkumarpilli/RekuAI.git && cd RekuAI

2. Setup virtual environment

uv venv .venv/bin/activate

3. Install dependencies

uv sync uv build

4. Pull data from GCP via DVC

dvc pull

5. Run the app locally

uv run application.py

📦 Dockerized App

bash

docker build -t rekuai-app . docker run -p 5000:5000 rekuai-app

⚙️ CI/CD with Jenkins DinD

Jenkins is configured in Docker-in-Docker (DinD) mode to enable container builds.

Triggers on push to main branch.

Runs tests, rebuilds image, pushes to GCR, and deploys to GKE via kubectl.

📄 See Jenkinsfile for the full pipeline.

☁️ Google Cloud Deployment

GCS – Stores and serves anime datasets.

GKE – Hosts the recommendation engine as a scalable pod.

DockerHub/GCR – Stores container images.

Ingress – Exposes service endpoint for web app and API.

🌐 Live Endpoint: Preview

🖥️ User Interface

Built with FastAPI and served via Kubernetes.

Simple search & recommend UI

Returns a list of top anime recommendations for a given user or title.

🎯 Sample Output

json
{
  "input": "Naruto",
  "recommended": [
    "Bleach",
    "Fullmetal Alchemist",
    "One Piece",
    "Attack on Titan",
    "Demon Slayer"
  ]
}

👥 Contributors Yogesh Kumar Pilli

💌 Contact 📧 Email: pilliyogeshkumar@gmail.com 🌐 GitHub: @yogeshkumarpilli